Extracting mathematical semantics from LATEX documents

Abstract

We report on a project to use SGLR parsing and term rewriting with ELAN4 to extract the semantics of mathematical formulas from a LATEX document and representing them in MathML. The LATEX document we used is part of the Digital Library of Mathematical Functions (DLMF) project of the US National Institute of Standards and Technology (NIST) and obeys project-specific conventions, which contains macros for mathematical constructions, among them 200 predefined macros for special functions, the subject matter of the project. The SGLR parser can parse general context-free languages, which suffices to extract the structure of mathematical formulas from calculus that are written in the usual mathematical style, with most parentheses and multiplication signs omitted. The parse tree is then rewritten into a more concise and uniform internal syntax that is used as the base for extracting MathML or other semantical information. © Springer-Verlag Berlin Heidelberg 2003.